dissenting.
The majority holds that a document in a government file drawer can violate a constitutional liberty interest in reputation and future employment. It holds in particular that the Newport News Police Department may have deprived former probationary employee Christopher Sciolino of “life, liberty, or property” by explaining in a letter to Sciolino himself that Sciolino’s employment was being terminated due to alleged misconduct and then placing the letter in its files. So mistrustful is the majority of state and local governments that it extends the Fourteenth Amendment to forbid actions that unpredictable courts predict create a “likelihood” of future harm. In this respect, as in others, the majority fails to grapple with the many practical problems of its holding.
My fine colleagues also contend that Newport News has made “charge[s] against [Sciolino] that might seriously damage his standing and associations in his community.” Maj. Op. at 652 (quoting Roth, 408 U.S. at 573, 92 S.Ct. 2701). This implies that the city has gone public in some fashion. See Bishop v. Wood, 426 U.S. 341, 348, 96 S.Ct. 2074, 48 L.Ed.2d 684 (1976) (requiring that information concerning discharge be “made public” to infringe constitutional interest). Newport News has done nothing of the sort. It has done nothing but keep its own record, which it must do if it is not to act arbitrarily and if it is to protect itself from future litigation. Thus to say that Newport News has somehow jeopardized the plaintiffs standing in the community is a complete mischaracterization of the city’s conduct. There is no contention — much less *654evidence — that Newport News made the letter public or sent the letter to any future employer. In fact, the majority’s “likelihood” standard concedes as much. Newport News has caused the plaintiff no reputational harm. It has acted in a way that would avoid causing the plaintiff repu-tational harm, so it has not deprived anyone of any Fourteenth Amendment interest.
I would not create a constitutional cause of action out of record-keeping functions. I respectfully dissent from the majority’s departure from constitutional text and Supreme Court precedent, a departure that extends judge-made law over routine aspects of the employment relationship in every state and local government in our circuit.
I.
There is at least one thing upon which the majority and I can agree: The circuits are split as to whether stigmatizing allegations can deprive a former government employee of a constitutional liberty interest before the allegations are disseminated to prospective employers or others.
The First and Seventh Circuits have written that such statements can only deprive a former government employee of a liberty interest when the allegations are publicly disseminated. See Johnson v. Martin, 943 F.2d 15, 16-17 (7th Cir.1991); Burton v. Town of Littleton, 426 F.3d 9, 17 (1st Cir.2005); Wojcik v. Mass. State Lottery Comm’n, 300 F.3d 92, 103 (1st Cir.2002). The Third Circuit has suggested the same, in my view, by stating that a plaintiff “must produce evidence that the reason for his termination was made public by the city” and rejecting a plaintiffs argument that he had been deprived of liberty because future disclosure was likely. Copeland v. Philadelphia Police Dep’t, 840 F.2d 1139, 1148 (3d Cir.1988).
The Second and Tenth Circuits have written, in contrast, that stigmatizing allegations can deprive a person of a liberty interest even if the allegations have not been publicly disclosed. The Second Circuit held in Brandt v. Board of Cooperative Educational Services that a liberty interest is implicated “where the stigmatizing charges are placed in the discharged employee’s personnel file and are likely to be disclosed to prospective employers.” 820 F.2d 41, 45 (2d Cir.1987) (emphasis added). The Tenth Circuit also recognized a claim based upon statements that the plaintiff did not allege went beyond a government employer’s own offices. Bailey v. Kirk, 111 F.2d 567, 580 n. 18 (10th Cir.1985).
Other circuits do not fit neatly into these two camps. The Eighth Circuit has suggested that a claim would lie if a plaintiff established that a record “would be available to prospective employers,” Clark v. Mann, 562 F.2d 1104, 1116 (8th Cir.1977) (emphasis added), but suggested in another case that information must be disclosed before it could give rise to a cause of action, Merritt v. Reed, 120 F.3d 124, 126 (8th Cir.1997).
The Fifth Circuit’s cases also seem to be in some internal tension. The court appears to have sanctioned claims based upon the possibility that allegations would be disclosed in the future, writing that an ex-employee could establish deprivation of a liberty interest by showing “that his employer has made or is likely to make the allegedly stigmatizing charges public in any official or intentional manner.” In re Selcraig, 705 F.2d 789, 796 n. 6 (5th Cir.1983) (emphasis added; internal quotations omitted). Yet a subsequent court rejected a plaintiffs likelihood-of-dissemination theory, equating it with an argument that “the mere presence” of stigma*655tizing allegations in a personnel file was sufficient to create a triable issue of fact. Hughes v. City of Garland, 204 F.3d 223, 228 (5th Cir.2000).
Finally, other circuits have recognized liberty claims when authorities lacked the power to keep statements confidential because personnel files were publicly available under state law, but have not addressed whether predictions of dissemination would suffice in the absence of such statutes. See Cox v. Roskelley, 359 F.3d 1105, 1110-12 (9th Cir.2004); Buxton v. City of Plant City, 871 F.2d 1037, 1042-46 (11th Cir.1989).
Thus to say there is a circuit split is at once true and not indicative of the full extent of the problem. Whether a liberty interest is infringed by a letter in a file drawer has generated answers with shades and permutations that mock the clarity law must provide for human conduct.
II.
A.
While both the majority and I recognize the circuits’ division, we diverge after that. Our disagreements run first to the treatment of constitutional text and structure. The majority appears to regard the conflicting circuit court rulings as a menu from which to select the standard of liability that it believes would amount to the best policy. It concludes that a terminated public employee may raise a Fourteenth Amendment claim based upon stigmatizing information in government personnel files even if the charges have never been made public,1 so long as the plaintiff alleges “a likelihood that prospective employers ... or the public at large will inspect the file” at some point in the future. Maj. Op. at 650. I reject this conclusion because far from authorizing the creation of such an interest, constitutional text and Supreme Court precedent foreclose both the holding in this case and the approach used to reach it.
While the rights of public employees may be protected in many different ways, the text of the Fourteenth Amendment does not permit the recognition of a liberty interest in not having stigmatizing information on file. By declaring that no state shall “deprive any person of life, liberty, or property, without due process of law,” the amendment permits judicial oversight of government procedures relating only to actions that “deprive” persons of interests within the named categories of “life, liberty, or property” — categories that are “not infinite.” Bd. of Regents of State Colls. v. Roth, 408 U.S. 564, 570, 92 S.Ct. 2701, 33 L.Ed.2d 548 (1972).
The Fourteenth Amendment did reorder the relationship of the federal and state governments in fundamental respects, but it did not displace state law as a residual protector of many interests the Fourteenth Amendment did not enumerate. Holdings that would “mandat[e] judicial oversight of communications between and among government employees and their superiors in the course of official business” can lead to “permanent judicial intervention in the conduct of governmental operations to a degree inconsistent with sound principles of federalism and separation of powers.” Garcetti v. Ceballos, — U.S. -, 126 S.Ct. 1951, 1961, 164 L.Ed.2d 689 (2006). In light of the amendment’s text and the federal structure, the Supreme Court has *656held that loss of reputation, government employment, and diminished professional prospects rise to the level of Fourteenth Amendment deprivations only under limited conditions that the majority disregards.
To begin with, it is undisputed that as a probationary employee the plaintiff has no Fourteenth Amendment property interest in a government job itself. As a result, all agree, he had neither a right to a hearing before the termination of his employment nor a right to know the reasons for his dismissal. Bishop, 426 U.S. at 348, 96 S.Ct. 2074. The loss of a government job and a filed record detailing the reasons for that loss are obviously no small matter. But the Constitution cannot possibly remedy every wrong experienced in life or for that matter every wrong experienced at the hand of some public entity. So the question must be where constitutional law properly leaves off and where statutory and common law remedies kick in. The question must be asked: the Fourteenth Amendment cannot be read to displace state remedial mechanisms by rendering the Constitution “a font of tort law to be superimposed upon whatever systems may already be administered by the States.” Paul v. Davis, 424 U.S. 693, 701, 96 S.Ct. 1155, 47 L.Ed.2d 405 (1976).
While it is constitutionally safeguarded in some instances, “reputation” is neither listed nor enumerated in the Fourteenth Amendment as a protected interest or right. There is no basis to elevate repu-tational harm above others into a constitutional interest because “[t]he words ‘liberty’ and ‘property’ as used in the Fourteenth Amendment do not in terms single out reputation as a candidate for special protection over and above other interests that may be protected by state law.” Id. A claim based upon stigmatizing government statements can therefore arise neither alone nor in the abstract. Rather, because reputation is indistinguishable from a multitude of personal, dignitary, and proprietary interests that state courts and legislatures may deem worthy of protection, the cases establish that neither the harms from the loss of at-will employment nor the harms from stigmatizing information rise by themselves to the level of constitutional harm. Id.; Roth, 408 U.S. at 575, 92 S.Ct. 2701. It is only when termination is combined with stigmatizing disclosures that the resulting harm is of constitutional magnitude.2
The majority’s conception of Fourteenth Amendment deprivation is as flawed as its discussion of Fourteenth Amendment interests. The plaintiff has not yet been deprived of an interest in his reputation, because the reasons for his dismissal evidently remain in a government file. Moreover, the liberty “to engage in any of the common occupations of life” posited in Meyer v. Nebraska, 262 U.S. 390, 399, 43 S.Ct. 625, 67 L.Ed. 1042 (1923), has not been infringed where the employer has neither made any information public nor shared information with a future employer. In short, the plaintiff has suffered no deprivation of reputation or of his ability to pursue other jobs.
*657The majority nevertheless holds that the Constitution can be violated if it finds “a likelihood” of future disclosure and reputa-tional harm. Maj. Op. at 650. The Constitution does not give courts the power to predict future harms and declare the Constitution violated before the possible harms transpire. The text makes it actionable for states to “deprive” a person of “life, liberty, or property,” but not to create a “likelihood” of such a deprivation. Whether the majority addresses Fourteenth Amendment “interests” or “deprivations,” it has done violence to our constitutional text and reordered our constitutional structure. The upshot is to federalize myriad aspects of the local employment relationship without any pretense of democratic sanction.
B.
If there was room for doubt about this matter, the Supreme Court eliminated it in Bishop v. Wood, 426 U.S. 341, 96 S.Ct. 2074, 48 L.Ed.2d 684 (1976). Bishop directly took up the question of when allegations of employee misconduct in connection with a termination implicate a Fourteenth Amendment interest. The plaintiff alleged that a police department deprived him of liberty when it terminated his employment and provided him with false and stigmatizing reasons for his discharge that might “severely damage his reputation in the community.” Id. at 347, 96 S.Ct. 2074. The Supreme Court found that in the absence of disclosure, the employer had caused no constitutional harm. Since the explanation of the plaintiffs discharge “was not made public ” before the plaintiff disclosed the allegations through his own lawsuit, the Court wrote that the explanation “cannot properly form the basis for a claim that petitioner’s interest in his ‘good name, reputation, honor, or integrity’ was thereby impaired.” Id. at 348, 96 S.Ct. 2074 (emphasis added; footnote omitted). The Fourteenth Amendment was not implicated, the Court concluded, by “the discharge of a public employee whose position is terminable at the will of the employer when there is no public disclosure of the reasons for the discharge.” Id. (emphasis added). To like effect is Codd v. Velger, which summarized the cases on constitutionally mandated hearings by noting, “Only if the employer creates and disseminates a false and defamatory impression about the employee in connection with his termination is such a hearing required.” 429 U.S. 624, 628, 97 S.Ct. 882, 51 L.Ed.2d 92 (1977) (emphasis added).
The majority acknowledges the binding precedent on this matter, as it must, but seems inexplicably puzzled by the Supreme Court’s plain language. The Supreme Court has used the words “made public” and “public disclosure” to describe the threshold of constitutional harm, but the majority writes that Bishop did not “resolve the question” of what a plaintiff must allege to meet this requirement. Maj. Op. at 648; see also Maj. Op. at 649. This observation wishes for ambiguity where there is none. To “make public” means “to cause to become known generally” just as “disclosure” entails “exposure” or “revelation,” and the majority cites no alternative definition or use that calls into doubt the meanings familiar to school children. The Random House Dictionary of the English Language 562, 1562-63 (2d ed.1987). As the Seventh Circuit has written, adopting a “likelihood of disclosure” standard for liability based upon stigmatizing personnel records requires “[djefining ‘public disclosure’ in a way which encompasses ‘no public disclosure’ ” — “an exercise we choose not to embrace.” Martin, 943 F.2d at 17.
Perhaps in recognition of these problems, the majority casts about for other *658possible injuries on which it can hang judicial intervention, seizing upon a potential indirect consequence of virtually any employee’s termination. It writes that a former employee might fear that prospective employers would make inquiries about his past job performance and might limit his job search as a result. See Maj. Op. at 649. Although these possible chilling effects are the only harm that a plaintiff could be said to have suffered from a letter in a file, my colleagues would evidently not require a plaintiff to plead or prove that he suffered even this indirect injury.
Moreover, this attenuated consequence cannot rise to the level of a constitutional deprivation, because injuries that are more serious and direct do not warrant constitu-tionalization. Roth, Bishop, and Paul did not reject Fourteenth Amendment claims because of blindness to the negative professional or reputational consequences that might flow from employees’ terminations or the records pertaining to them — the cases explicitly recognized such consequences. See Roth, 408 U.S. at 574 n. 13, 92 S.Ct. 2701; Paul, 424 U.S. at 701, 96 S.Ct. 1155. They nevertheless refused to find a constitutional interest because the text and structure of the Fourteenth Amendment set too high a bar.
III.
A.
Even if judges were somehow set free to prescribe policy unconstrained by text or precedent, a lively debate might ensue over whether the majority view is suitably wise. My friends make a case that it is, but as with most matters of policy, there exist two sides.
First, slighting the principle that “[t]he federal court is not the appropriate forum in which to review the multitude of personnel decisions that are made daily by public agencies,” Bishop, 426 U.S. at 349, 96 S.Ct. 2074 (footnote omitted), widens the gap between public and private employees. The majority observes that a former government employee who feared the release of stigmatizing allegations might feel obligated to “choose between finding future employment and protecting his reputation by not applying for jobs.” Maj. Op. at 649.
That is all well and good, but these are the precise harms that every person terminated from employment may experience. While the prohibitions on state action in the Fourteenth Amendment will always necessitate some differences in treatment between public and private employees, the majority’s standard as to likelihood of harms drives a wedge between these two groups of workers to an extent that many lawmakers might find both divisive and unpalatable. Whether widening the gap in our country between private and public employees is desirable is at least an arguable point, as is whether the majority’s approach leaves two groups with very different remedies for comparable harms without compelling reason.
Second, the broad power that the majority claims over personnel matters invades an area in which state and local officials are often better equipped than judges to make decisions. Whether benefits of accuracy and fairness justify the costs of particular procedural safeguards is not a question that can be answered in the abstract and for all time. Leaving the analysis to democratically responsive local authorities where the Constitution’s text does not permit judicial involvement makes it possible to strike balances suited to local needs and facilitates adaptation to changing circumstances and new insights.
For example, in permitting liability based upon a letter in a personnel file, the majority may make government processes
*659less fair and more arbitrary by “penalizing] forthright and truthful communication between employer and employee.” Bishop, 426 U.S. at 349, 96 S.Ct. 2074. The majority makes it costlier to institute systems of progressive discipline that require documentation of misdeeds and take into account past conduct in imposing penalties, even though such systems may benefit both employers and employees. Furthermore, if a government employee is terminated, he may ask for a reason and his employer may consider it to be in the interests of fairness to provide an explanation even when there is no legal obligation to do so. But when courts open governments to liability because they provide such reasons, government employers may simply stop giving them: “where there was no constitutional requirement for the state to do anything,” Judge Friendly has written, “the state would merely opt to give no reasons and the employee would lose the benefit of knowing what might profit him in the future.” Russell v. Hodges, 470 F.2d 212, 217 (2d Cir.1972). Tenured employees, of course, are entitled not simply to know the reasons for their dismissal but to challenge them through hearings, see, e.g., Roth, 408 U.S. at 576-77, 92 S.Ct. 2701, and it seems anomalous to require the provision of reasons to tenured employees while discouraging the provision of reasons to untenured ones.
Even with respect to untenured employees, some acts of documentation may not be wholly voluntary, and as a result, the majority’s new threat of liability may place governments between a rock and a hard place. Procedural and substantive limits on employee discipline are common subjects of collective bargaining, and union agreements may therefore require documentation of the reasons for dismissals. See Ann C. Hodges, The Interplay of Civil Semce Law and Collective Bargaining Law in Public Sector Employee Discipline
Cases, 32 B.C. L.Rev. 95, 104-07 (1990). Governments may also record these reasons in order to defend themselves in the event that a former employee claims a dismissal was based upon race, sex, or some other impermissible consideration. Under Title VII of the Civil Rights Act of 1964, for example, once an employee presents a prima facie case of unlawful discrimination — a showing that is “not onerous” to make — the defendant “must clearly set forth, through the introduction of admissible evidence” the “legitimate, nondiscriminatory reason” for its challenged action. Tex. Dep’t of Cmty. Affairs v. Bur dine, 450 U.S. 248, 253, 255, 254, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981). This leaves governments with little choice but to document the reasons for major personnel actions, and it is odd, to say the least, to make such records necessary and in the next breath open governments up to liability on the basis of the same documents.
B.
The practical problems with the majority’s approach do not end there. It is no answer to say state and local governments need only avoid constitutional infractions to avoid being found in violation of the Fourteenth Amendment because under the majority’s vague standard, this is more easily said than done. The majority imposes an amorphous overlay upon an area of the law where balancing tests already leave state and local governments uncertain about the nature of their obligations. Government employers already face a difficult predictive calculus in establishing procedures to govern actions that indisputably deprive citizens of liberty or property, since the procedures required depend upon judicial balancing of
[f]irst, the private interest that will be affected by the official action; second, the risk of erroneous deprivation of such *660interest through the procedures used, and the probable value, if any, of additional or substitute procedural safeguards; and finally, the Government’s interest, including the function involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail.
Mathews v. Eldridge, 424 U.S. 319, 335, 96 S.Ct. 893, 47 L.Ed.2d 18 (1976).
The standard of liability is vague in a second important respect. Employers face additional uncertainty over whether countless personnel documents generated in the course of government operations “imply the existence of serious character defects,” and could therefore implicate a constitutional interest if disclosed. Robertson v. Rogers, 679 F.2d 1090, 1092 (4th Cir.1982). Adverse personnel actions are by definition taken on the basis of unfavorable assessments, and the majority provides little guidance for gauging which of these assessments are sufficiently unflattering to qualify as actionable.
And the majority now adds a third level of uncertainty: Government employers will have to predict whether a court would find a “likelihood” that even confidential documents would be disclosed in the future. This vague standard imposes enormous costs even when government conduct is found to be blameless, because uncertainty over its application will give rise to litigation in the absence of harm.
This case perfectly illustrates the degree to which the majority’s standard will put governments at the mercy of litigants no matter how careful the governments’ conduct. Newport News must have imagined it would not be drawn into court under a likelihood-of-dissemination standard: The city concluded that disclosure of the plaintiffs personnel files would be not merely against its own policies but prohibited by law under all but specified circumstances.
Under Virginia’s Government Data Collection and Dissemination Practices Act, a government agency may “disseminate only that personal information ... necessary to accomplish a proper purpose of the agency.” Va.Code Ann. § 2.2-3803(1). City policy accordingly forbids disclosure of the reasons for an employee’s termination except when failure to disclose the reasons to a requesting government agency could place the requestor in a high liability situation. Indeed, the district court agreed that state law would prohibit disclosure of Sciolino’s personnel files. Yet neither of these facts saves the city from litigation— and perhaps liability — under the majority’s “likelihood” view. See Maj. Op. at 647 n. 3. It is hard to see how almost any statement will not be fodder for litigation under this approach.
Nor can a city avert litigation by offering generous procedural rights because as this case shows, a plaintiff can still enmesh the government in litigation by alleging that yet more safeguards are required. In this case, before Police Chief Mook dismissed Sciolino from his probationary position and wrote him a letter documenting findings of misconduct, Mook met with Sciolino to explain the charges and offer the opportunity to respond. Although the majority concedes that the meeting “may have afforded Sciolino all the process to which he would be due,” it keeps the city in court in order to keep its own options open, writing that “[t]he record in this case is not sufficiently developed” to assess the adequacy of the procedural safeguards “at this early stage.” See Maj. Op. at 646 n. 1. As this case illustrates, there is little that state and local governments will be able to do to avoid litigation over even meritless claims under the unascertainable standards the majority imposes.
Finally, it bears repetition that this whole area reflects a tension between two *661competing and legitimate interests. On the one hand, there is the interest of the employee in avoiding any negative consequences from stigmatizing accounts of his job performance. On the other, there is the employer’s interest in documenting serious employee misconduct and in holding open the prospect of honest and accurate communications between employers about prospective hires who have serious character defects or have even engaged in criminal misdeeds. Where there has been a deprivation of a constitutional interest, Mathews requires that courts strike the balance between these considerations. Where, as here, there has been no loss of reputation and indeed no deprivation of any interest rising to constitutional magnitude, legislatures and state governments remain free to balance the employee’s interests against the need for maintaining records that will facilitate accurate decision-making. Where the most the public employer has done is put a letter in a file, the balancing of the interests falls to legislatures and not to the courts.
C.
My colleagues evidently impose this vague and burdensome standard for constitutional liability because they are troubled that if municipal liability could arise only if allegations were disseminated, courts would be restricted in their ability to rework government procedures until after reputational harm occurred. The majority finds this limitation too great to countenance: “If an allegation of actual dissemination were required, the information would have already been communicated to a potential employer, the employee’s job opportunities foreclosed, and his reputation damaged before any possibility for a [judicially ordered] name-clearing hearing.” Maj. Op. at 650. In addition, a former employee might have to seek discovery in order to be certain that he had a cause of action, because he might not know before filing suit whether personnel information had been communicated to members of the public. Maj. Op. at 650.
This reasoning is flawed in several ways. It utilizes a dangerous and forbidden mode of judicial interpretation, reasoning from the remedy that judges wish to provide back to the creation of a constitutional wrong. This transforms the Due Process Clause from a protection of a named set of rights into a roving commission to rework state and local governments in the manner that judges deem optimal by finding “rights” when necessary. The Supreme Court has confronted and rejected this judicial approach, writing repeatedly that under the Due Process Clause, “The categories of substance and procedure are distinct.” See Cleveland Bd. of Ed. v. Loudermill, 470 U.S. 532, 541, 105 S.Ct. 1487, 84 L.Ed.2d 494 (1985). The first step in a due process inquiry must be to determine “whether the nature of the interest is one within the contemplation of the ‘liberty or property’ language of the Fourteenth Amendment,” with courts proceeding to assess the adequacy of process only if so. Morrissey v. Brewer, 408 U.S. 471, 481, 92 S.Ct. 2593, 33 L.Ed.2d 484 (1972); see also Loudermill, 470 U.S. at 541, 105 S.Ct. 1487. Roth unambiguously rejected a remedy-first approach, finding that a district court erred when it engaged in such reasoning. See 408 U.S. at 569-71, 92 S.Ct. 2701.
Moreover, the limitation on judicial power that the majority finds so troubling is a premise of our constitutional system that has served us well for centuries. The majority writes as though it would be remarkable and regrettable that the plaintiff would have to wait until he suffered repu-tational harm before claiming that his constitutional rights had been violated, but *662under the conception of the judicial role embedded in our Constitution, a case must involve actual or imminent injury to even be justiciable. Compare Maj. Op. at 649 (“Requiring a plaintiff to wait until he actually loses some job opportunities would place him between the devil and the deep blue sea”) (internal quotations omitted) with, e.g., City of Los Angeles v. Lyons, 461 U.S. 95, 101-02, 103 S.Ct. 1660, 75 L.Ed.2d 675 (1983) (“The plaintiff must show that he has sustained or is immediately in danger of sustaining some direct injury as a result of the challenged official conduct and the injury or threat of injury must be both real and immediate, not conjectural or hypothetical”) (internal quotations omitted). I am not certain that even the requirements of standing are satisfied in this case: A plaintiff may have a cause of action based upon a “likelihood” of future disclosure under the majority’s theory, even if the plaintiff is not in immediate danger of reputational harm or other constitutionally cognizable injury.3 Proceedings in the absence of injury exceed our judicial authority and add unnecessary litigation costs to state and local government budgets.
Nor is it at all remarkable or unique that a potential plaintiff may not have all facts surrounding a possible violation of his constitutional rights before filing suit. While the majority evidently finds unacceptable the Supreme Court’s requirement that statements be “made public” in order to give rise to liability because it believes that this requirement makes due process protections “virtually impossible to enforce,” see Maj. Op. at 650, plaintiffs often do not have all the knowledge necessary to prove their causes of action before filing suit. If they did, there would be no need for the many provisions in the Federal Rules governing the conduct of discovery. See, e.g., Fed.R.Civ.P. 11(b)(3), 26-37. The majority’s standard would not obviate the need for this discovery. To the contrary, plaintiffs will be required to undertake more extensive discovery and seek the testimony of a greater array of officials now that their cases turn not upon the simple question of whether files have been released but upon the speculative question of whether there exists a likelihood that they would be at some time in the future.
The majority need not fear that state and local governments will run riot in the absence of its novel constitutional doctrine. The elected branches have compelling reasons to safeguard citizens’ reputational interests absent the strong medicine of judicial oversight of their internal workings. They are accountable to constituents and often answerable to public employee unions; indeed, they have been providing mechanisms to vindicate reputational interests through causes of action such as defamation since long before the Constitution was invoked for this purpose. Many governments have added to these protections through civil service systems, which “typically restrict the public employer’s discretion to discharge and to impose other serious discipline, such as demotion and suspension,” and often allow dismissed employees to challenge the reasons for a dismissal through an appeals process. Hodges, supra, at 103.
Indeed, a government’s own effective functioning is served by procedures to ensure that employee morale remains high and that good employees are not lost due to dismissal based upon inaccurate reports. An “actual dissemination” standard for constitutional liability itself creates in*663centives to institute procedural protections before the fact. Even when government actors have not made personnel reports public, agencies that wish to reserve the right to disseminate in the future would be well advised to hold hearings when all parties relevant to allegations of misconduct are available and their memories are fresh. Under an actual dissemination standard, however, the judgment about future likelihood would be reserved to the actor most capable of making it — the entity that might disseminate the allegedly stigmatizing statements — and courts would not be drawn into wholly hypothetical disputes about the precise level of name-clearing process required in light of the injurious effects of disclosures that may never occur.4
IY.
The Due Process Clause preserves essential freedoms, but it may also tempt judges to overrun their role in order to impose their own conceptions of justice on state and local governments. There is room for constitutional standards in this area, as the Supreme Court has made clear. But placing states and localities at constitutional risk for performing record-keeping functions goes much too far. By over-constitutionalizing public employee relationships, as the majority does here, courts diminish the role of all the other participants in this field and assume that neither employers, nor employees and their representatives, nor state law rights and remedies can fairly be trusted to reach just resolutions. The Supreme Court has sought to avoid the danger of over-consti-tutionalizing in its own procedural due process cases. When Roth rejected a professor’s claim that he should be provided a hearing to contest his termination, the Supreme Court explained,
Our analysis of the respondent’s constitutional rights in this case in no way indicates a view that an opportunity for a hearing or a statement of reasons for nonretention would, or would not, be appropriate or wise in public colleges and universities. For it is a written Constitution that we apply. Our role is confined to interpretation of that Constitution.
408 U.S. at 578-79, 92 S.Ct. 2701 (footnoted omitted).
The majority fails to heed those words. It creates a nebulous new right whose contours assure nothing but the presence of continued litigation and the permanence of judicial oversight. Such are the costs of neglect of text and structure. Whether the majority’s views are wise I am content to leave to others, but for reasons stated, I would affirm the district court.
. The plaintiff may have made the allegations against him publicly available by filing his lawsuit, but the Supreme Court has held that self-generated exposure in litigation does not give rise to a constitutional claim. Bishop v. Wood, 426 U.S. 341, 348-49, 96 S.Ct. 2074, 48 L.Ed.2d 684 (1976).
. The majority contends that a failure to adopt its likelihood standard would strip the Due Process Clause of its historic function in providing pre-deprivation hearings. Maj. Op. 653 n. 9. A pre-deprivation hearing, however, is mandated only when a public entity is proposing to deprive an individual of a constitutionally protected right or interest, and, even in such circumstances, a pre-deprivation hearing may not always be required. See, e.g., Parratt v. Taylor, 451 U.S. 527, 540-41, 101 S.Ct. 1908, 68 L.Ed.2d 420 (1981). Here, the City of Newport News is not proposing such an action, and it is improper to subject a public entity to suit purely on the basis of a record-keeping function.
. The plaintiff's dismissal itself cannot be the source of his claimed injury because it is undisputed that an untenured probationary employee has no constitutional interest in his government job. See supra at 65 6.
. Even under the majority's likelihood-of-dissemination standard, I would affirm the disposition below. Once judgment had issued, the district court did not abuse its discretion in denying the plaintiff's motions under Rule 59(e) and Rule 15(a) to reopen his case and to amend his complaint for a second time. While amendment is permitted when, prior to discovery, a plaintiff "merely adds an additional theory of recovery to the facts already pled,” Laber v. Harvey, 438 F.3d 404, 427 (4th Cir.2006) (en banc), or when we overrule a precedent that governed the plaintiff’s claim at the time of filing, id. at 428, neither circumstance is present here. The district court therefore did not abuse its discretion by giving weight to the interests of finality shared by the defendants and the judicial system itself.